Path: news.mathworks.com!not-for-mail
From: <HIDDEN>
Newsgroups: comp.soft-sys.matlab
Subject: Re: Parseval's Theorem
Date: Thu, 12 Mar 2009 21:45:19 +0000 (UTC)
Organization: Xoran Technologies
Lines: 23
Message-ID: <gpbvpf$ivd$1@fred.mathworks.com>
References: <gkahee$1t3$1@fred.mathworks.com> <1d0a6b1e-94f8-43e9-811d-aedfab2fe755@n2g2000vbl.googlegroups.com>
Reply-To: <HIDDEN>
NNTP-Posting-Host: webapp-03-blr.mathworks.com
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-Trace: fred.mathworks.com 1236894319 19437 172.30.248.38 (12 Mar 2009 21:45:19 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Thu, 12 Mar 2009 21:45:19 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 1440443
Xref: news.mathworks.com comp.soft-sys.matlab:524446

Just to make sure the dust settled on this issue, I'm putting some examples of how to orthonormalize MATLAB's FFT and IFFT functions (i.e. to make them consistent with Parseval's theorem)  

a=rand(10); 

A=fftn(a);  A=A/sqrt(numel(A));  

anew=ifftn(A); anew=anew*sqrt(numel(anew));

>> norm(a(:))^2, norm(A(:))^2, norm(anew(:))^2

ans =

   36.6115


ans =

   36.6115


ans =

   36.6115